Self-powered, flexible and room temperature operated solution processed hybrid metal halide p-type sensing element for efficient hydrogen detection

Hydrogen (H2) is a well-known reduction gas and for safety reasons is very important to be detected. The most common systems employed along its detection are metal oxide-based elements. However, the latter demand complex and expensive manufacturing techniques, while they also need high temperatures or UV light to operate effectively. In this work, we first report a solution processed hybrid mixed halide spin coated perovskite films that have been successfully applied as portable, flexible, self-powered, fast and sensitive hydrogen sensing elements, operating at room temperature. The minimum concentrations of H2 gas that could be detected was down to 10 ppm. This work provides a new pathway on gases interaction with perovskite materials, launches new questions that must be addressed regarding the sensing mechanisms involved due to the utilization of halide perovskite sensing elements while also demonstrates the potential that these materials have on beyond solar cell applications.

the figures of merit of an ideal sensing element, despite their appealing characteristics: small size, high sensitivity, high repeatability and simplicity to use. Often, they suffer from some serious drawbacks such as (a) high operation temperatures (operation at room temperature has shown weak response to low concentration H2 or have long response-recovery time.
Moreover operating at high temperatures is very risky especially when the targeted gas is an explosive gas such as H2); (b) they need a pre-treatment with UV light in order to become conductive; and (c) their fabrication is complicated and expensive (e.g. rf sputtering, dc sputtering, high vacuum facilities, pulsed laser deposition [16][17][18][19][20][21]. Hybrid metal halide perovskite materials are the new star materials in the solar cell technological field and not only. They are of the type ABX3 where, A is an organic cation usually methylammonium (MA + ), formamidinium (FA + ) etc. or an inorganic cation such as Cs + , Rb + and K + , B is a metal such as Pb 2+ , Sn 2+ , while the anion X is a halogen such as I -, Br -, Clor a mixture thereof. The very first report on semiconductor halide perovskite was via observing photoconductivity in the all inorganic CsPbX3 systems at late 1950'by Moller [22]. Later on, perovskite was used as absorbing material in photovoltaic applications by Kojima et al. in 2009 [23]. In the last decade, the efficiency of perovskite based solar cells have been skyrocketed from 3.8% to higher than 25% [24]. The hybrid lead halide perovskites, such as the mixed halide one studied here (CH3NH3PbI3−xClx), have recently attracted the scientific communities' attention as one of the most promising solar light energy harvesting materials due to their direct bang gap, long diffusion charge carrier lengths, large absorption coefficients, long carrier lifetime and large carrier mobility [25][26][27]. The impressive impact of perovskite in solar cell technology, have attracted an intensive research interests towards the applications of perovskites beyond solar cells such as in lasers [28] in light emitting diodes [29,30] and photodetectors [31,32]. However, one of the key issues to be solved in order to boost their commercialization is related to their stability. Halide perovskites are very sensitive to polar gases and vapours, as, exposure to such elements deteriorates substantially and very fast the perovskite devices' performance. The conversion of this disadvantage to an advantage reflects the introduction of perovskite-based gas sensing elements. The sensitivity of perovskite to environmental gases is an opportunity to convert a drawback to an opportunity [33]. Recently, halide perovskites have been explored and demonstrate their competence over the existed technologies, as potential sensing elements from gas molecules to X-ray photons [34][35][36][37][38]. Increasingly new publications regarding inorganic perovskites applications in H2 sensing, have started to appear [39,40] (Figure 1) showing the potential of this family of materials towards hydrogen sensing. However, all reported inorganic perovskite materials need to operate at very high temperatures (of few hundred Celsius) in order to function as hydrogen sensors. In this paper, we introduce for the first time a solution processed hybrid to other already demonstrated in hydrogen sensing materials such as metal oxides and metals [4]. More particular compared to metal oxides (e.g. SnO2, ZnO, TiO2, FeO, Fe2O3, NiO, Ga2O3, In2O3, Sb2O5, MoO3 and WO3) that request high temperatures to operate as hydrogen sensing elements, the demonstrated hybrid perovskite film operate in room temperature reducing a lot the consuming power. To highlight that the performance of both systems regarding the reaction times (of the order of few seconds) and minimum detection limits are similar (of the order of 10 ppm). Regarding their advantages compared to metallic resistors (e.g. palladium & platinum metals) (a) request simpler deposition techniques (e.g. spin coating) compared to more complicated and expensive techniques the metallic resistors request: vacuum evaporation, electrode position, sputtering and pulsed laser deposition; (b) do not suffer from mechanical degradation when exposed to hydrogen environment and thus do not request complicated metallic alloys to address the issue of the mechanical degradation. To add that hybrid perovskite semiconductors, show similar minimum detection limits as metallic resistors.
The studied sensing films demonstrated a p-type semiconductor behaviour (attributed to its stoichiometry) and thus under H2 gas (reducing gas) exposure their electrical resistance was enhanced upon H2 exposure. However, its pristine electrical properties were restored very fast (within few secs) after the removal of the H2 gas. This work is believed to set the framework for research of low temperature operated, efficient and of low cost new conductometric (resistance measurement), compatible with flexible electronic industry halide perovskite H2 sensing elements and systems. This is essential for various applications ranging from the energy sector to aerospace industry.

Methodology
To fabricate the CH3NH3PbI3−xClx precursor solution, methylammonium iodide (MAI, purchased by Dyesol) with lead (II) chloride (PbCl2, 99.999% purchased by Sigma Aldrich) were mixed in a molar ratio of 3:1 within a anhydrous N,Ndimethylformamide (DMF) solvent. The resulted solution (with 40 wt% concentration) was continuously stirred on a hot plate for 12 hrs at 70 °C. Afterwards, the precursor solution was cooled down at room temperature and deposited (approximately 20 μl filtered precursor solution) onto the sensing element electrodes template (purchased from DropSens). The latter and prior to the spreading of the solution precursor, was UV ozone cleaned to remove any hazards (e.g. organic contaminants) that can lower its hydrophilicity. The electrode substrate was glass made and on top of the glass two interdigitated platinum electrodes were pre-patterned. The distance between the electrodes was 5 μm. The deposited solution was spin-coated at 4000 rpm for 45 seconds. entire processing was conducted inside a nitrogen-filled glove box with O2 and H2O concentrations below 0.1 ppm. Prior to the testing of the sensing properties of the glass/Pt/CH3NH3PbI3−xClx sandwich elements, the latter were fully characterized as a function of their electronic, morphological, structural properties. Prior to the sensing testing the glass/Pt/CH3NH3PbI3−xClx system was exposed under simulated solar light (A.M 1.5G at 100mW/cm 2 ) in ambient conditions (~45% relative humidity). The film roughness was examined using Atomic Force Microscopy (AFM), by employing a Park XE-7 instrument in tapping mode. The total scan area was set to 50 μm x 50 μm and the scan rate was fixed at All the hydrogen sensing measurements occurred within a homemade gas test chamber under dark conditions in order to lower the rates of photoexcitation that contributes to dark current. Prior to the sensing test measurements that occurred, the electrical conductivity of our elements was tested. A potential difference of one Volt was applied across the two platinum electrodes of glass/Pt/CH3NH3PbI3−xClx systems and the induced generated current was measured using a Keithley 6517A multi-meter set-up. All these initial measurements were taken at room temperature and under a pressure of 1.6 mbar.
After the baseline of the conductivity of our sensing elements was measured, the CH3NH3PbI3−xClx sensors were exposed for five minutes to hydrogen gas at a constant flow of 500 sccm (standard cubic centimetres per minute), while the pressure in the chamber was kept constant at 120 mbar, leading to a decrease of current. The hydrogen concentrations that the detector was exposed were 100, 75, 50, 25, and 10 ppm. The sensing performance of our sensing elements could not be assessed for hydrogen concentrations higher than 100 ppm for safety reasons. The CH3NH3PbI3−xClx sensor was exposed for time intervals of five minutes at each concentration, while another five minutes was given to the sensor to relax to its steady-state conditions.

Structural, Morphological and Optical Analysis
A 300 nm thick CH3NH3PbI3-xClx film was fabricated onto a glass substrate with prepatterned electrodes; the interdigitated electrodes made of platinum had distance between them of 5 μm. It must be highlighted at this point that the majority carriers within the perovskite semiconducting film are subject to the stoichiometry of the PbI 2 /MA + I precursor ratio; in our case this ratio was less than one and the film demonstrated a p-type semiconducting behaviour [41]. The entire fabrication process was performed inside a glovebox under nitrogen atmosphere.
The film's surface morphology controls the number of the provided interaction pathways that the targeted gas molecules can interact on with the sensing element; increased roughness, facilitates the gas molecules adsorption by providing longer diffusion lengths within the active material. The surface film's morphology was revealed using Atomic Force Microscopy (AFM) measurements ( Figure 2). AFM patents were taken before ( Figure 2a) and after the exposure to H2 (Figure 2b). A striking change in the AFM images was observed, with the roughness to reduce almost to its half value after the exposure to 100 ppm of hydrogen gas. The reason of this morphological change is thought to relate to the interaction of H-molecules with the perovskite surface species; however, the exact cause needs further investigation before it can be attributed to a specific factor. Subsequently, a slight decrease in the measured current (of the order of few nA during the measurement cycles) was attributed to the resulted smoother surfaces; the latter provide shorter percolations paths to allow the H2 molecules to interact with the perovskite platform.  The film's crystal structure was studied using the X-ray diffraction (XRD) technique. The successful crystallization of the spin coated perovskite film is depicted in the acquired XRD image (Figure 4). The crystal directions of the first (110) and the second (220) crystallographic plane can be seen, at ~14.2 o and ~28.5 o respectively, confirming the, as expected, cubic perovskite phase [42]. No compositional changes were observed before and after the H2 to hydrogen gas molecules (at 100 ppm for more than one hour), showing the sensing potential and stability of the fabricated films. This is a clear indication that the interacted H2 molecules just adsorbed and de-adsorbed through the perovskite-based template. Figure 4: XRD images of the mixed halide perovskite films before and after its exposure to 100 ppm hydrogen gas. The results showed no structural changes after the interaction of the sensing element with the hydrogen. The XRD of the non-exposed halide perovskite film was taken on a glass substrate without electrodes

Sensing Properties of the CH3NH3PbI3-xClx thin films
The sensing functionality of the films on H2 was examined by electrical measurements at room temperature, carried out in a home-made set-up. We exposed the aforementioned perovskite films to various H2 gas concentrations up to the maximum of 100 ppm for safety reasons.
Interruption the H2 gas flow, restored the films' conductivity to their initial levels. The excellent electrical properties (long diffusion lengths, high charge mobility and excellent mobility to charge lifetime product), the interdigitated electrodes (Pt electrodes with distance of the order of 10 μm) setup and the application of a forward bias across the electrodes (0.5 and 1 V) allowed us to register (a) the current across the film (of the order of μΑ); and (b) its modulations when the film was exposed to H2 environment. The report of such high currents (in μΑ range) through the film is an indication of its excellent electrical properties as a result of the long grain's size and high carrier's diffusion length of the CH3NH3PbI3−xClx perovskite layer employed as sensing material. The demonstrated sensing elements were self-powered thus there was no the need of any external assistance such as UV irradiation or heating in high temperatures. H2 molecules operated as a reducing gas, and since the exhibited perovskite film is a p-type semiconductor, lowering of the current was observed, as expected, caused by the adsorption of the H2 molecules (the opposite behaviour than n-type semiconductors).  During the sensing measurement process, the modulation of the current flowing through the film was recorded, as the H2 gas in the synthetic air was inserted and stopped, following several cycles. The film demonstrated sensing reversibility as this is illustrated in Figure 6.
Interruption H2 gas flow lead to the recovery of the prior sensing element electrical resistance.
The reproducibility of the acquired results, supports the credibility of our sensing element performance and reliability. The sensing film demonstrated a reversible sensing behaviour under various hydrogen gas concentrations: 100 ppm, 75 ppm, 50 ppm, 25 ppm and 10 ppm.
As a result, the reducing gas causes an increase in the film resistance, reconfirming its p-type character. As expected, our perovskite sensing platform demonstrated different sensitivity at different targeted gas concentrations ( Figure 6). This, in combination with the abrupt increase of the resistance, was attributed to the H2 molecules adsorption into the perovskite surface.
Furthermore, as the need for low temperature gas sensors is becoming a necessity and moreover for safety reasons in the case of the highly flammable H2 gas, all the measurements were taken at room temperature although sacrificing part of the response signal obtained at elevated temperatures. All the measurements were recorded until the current through the sample had reached its lowest value. This process duration was lasted for approximately five minutes after which no further substantial changes were observed. The sensing measurement was repeated under each hydrogen gas concentration for five times (five cycles) - Figure 6, showing excellent reproducibility. The sensing ability was assessed with the measurement of (a) the sensitivity (S); and (b) the response & recovery times of the film. The sensitivity parameter is calculated using the following formula (1) [43] (%) = | − | 100% (1) where Iair denotes the electrical current of the sensor before the exposure into the H2 environment, Igas the resistance of the sensor after ten minutes to hydrogen. The response (tresp.) and the recovery (trecov.) times were calculated as the times take the measured current to reach the 10% of its maximum value under H2 exposure and 90% of its       concentrations. You can see in Figure 10 below, that the maximum exposure time was 10s. The absence of any phase transformation and impurity formation, upon exposure to a H2 atmosphere, shows the non-chemical interaction of the hydrogen with the halide perovskite platform. This is also supported from the reversibility of the electrical properties of the perovskite-based film exhibited after the removal of the H2 gas molecules from the chamber.
It is crystal clear that the H2 molecules are adsorbed through the porous of the perovskite film and bond loosely with its crystal template close to the surface; this can be interpreted by the small response and recovery times the film showed. The increase of the films' resistance hydrogen is desorbed, the surface of the material continues to adsorb oxygen from the environment to generate oxygen ions. At this time, the resistance of the material returns to the base value. However, the XRD measurements showed no compositional changes (the formation of water molecules should lead to a severe degradation and decomposition of the halide perovskite). This does not occur even though the films were exposed to H2 environment for more than an hour. This is an open question. We conclude that further studies and theoretical modelling should be done to support further the aforementioned operational mechanism for the H2 sensing 25 .
It is noted that after a number of successive cycles, the current did not fully return to its initial value. This is attributed possibly to residual adsorbed H2 molecules due to slight changes in the film's morphology as the AFM measurements revealed. The observation of no compositional change after the exposure to H2 molecules is attributed to the good crystallinity of the perovskite film that reflects film's good stability. Moreover, the sensing properties in successive hydrogen/synthetic air switching cycles of the mixed halide films was tested after three weeks of storage under vacuum within the measurement chamber demonstrating remarkable sensing element durability. The results were encouraging, with the sensing element to demonstrate a good stability (see Figure 10), as the sensitivity and the detected currents remained almost the same after three weeks of storage. The ability to distinguish different gases was also demonstrated by the fabricated films. The same film exposed to the H2 gas, was exposed to ozone molecules (Figure 11a  where similar electrode patterns with these films on rigid substrates have been printed. The film was exposed to H2 environment under (a) no bending; and (b) bending conditions. In Figure 12a-c is depicted the sensing performance under bending conditions (Fig. 12b); and the two extreme cases of no bending and bending close to 180 o (Fig. 12c); in both cases the modulation of current due to H2 interaction with the perovskite template could be detected and be measured. This is the first time that a report on the flexibility of the tested material as sensor is made and these results are very promising. The sensitivity recorded was comparable (around 7% under bending and exposed to 100 ppm of hydrogen molecules) with the one of the mixed halides films deposited onto glass substrates.

Conclusions
We report here for the first time, solution processed hybrid mixed halide perovskite films (CH3NH3PbI3-xClx) as gas sensor elements. The films were prepared and spin coated onto rigid (glass) and flexible (PET), prepatterned with interdigitated electrodes, substrates. All films operated as self-powered sensing elements for H2 at room temperature. Their sensing properties were based on modulations of its electrical resistance. H2 sensing measurements revealed very promising results regarding the potential of this material as H2 sensing element; the p-type semiconductor characterised by maximum sensitivities of the order of 5-7%, with very fast reaction times of the order of few seconds and the capability to be able to distinguish between two different tested gases (H2 and O3). The compatibility of the demonstrated sensing element with flexible substrates was also confirmed. Ageing measurements showed that the devices retained their sensing abilities even after three weeks of storage. However, despite the first promising results obtained, further work is planned be done in order to improve mixed halide perovskite film's sensing performance towards H2 gas molecules. First of all, the operational mechanism should be confirmed and studied further. Second, doped hybrid perovskite mixed halide films or nanocomposites with graphene-based materials that exhibit higher conductivities, should be tested. The impact of higher conductivities to the sensing performance must be linked. This is expected to enhance the acquired sensitivity.
Another idea, is to try to apply 2D solution processed perovskite materials since the lowering of dimension leads to higher surface to volume ratio and thus to the enhancement of all the figures of merit of a sensor: (a) higher sensitivities; (b) broader limit of detection; (c) higher stabilities. The better engineering of mixed halide perovskite probably will provide more tolerant to ion migration problem and will allow the application of higher applied voltages.
This is expected to elevate the sensitivity of the particular sensing elements. Stimulated emission Raman Spectroscopy principles will be employed as a tool to be able to distinguish the various gases in situ.
Based in the above, it may be concluded that the hybrid mixed halide perovskite films could be a promising self-powered p-type sensing material for reducing gases as H2 at room temperature.